Robust imputation method with context-aware voting ensemble model for management of water-quality data

Area of Science:

• Environmental Science
• Data Science
• Hydrology

Background:

• Water quality monitoring is vital for resource management.
• Missing data in water quality datasets can bias hydrological modeling and analysis.
• Existing imputation methods lack robustness across various missing data scenarios.

Purpose of the Study:

• To develop a robust imputation method for water quality data that addresses limitations of existing techniques.
• To create a context-aware voting-ensemble model for dynamic integration of imputation models.
• To improve the accuracy and reliability of water quality data imputation across diverse missingness scenarios.

Main Methods:

• Developed a context-aware voting-ensemble model with dynamic weighting.
• Identified attributes influencing missingness scenarios and imputation accuracy.
• Optimized model weights using regression to capture relationships between scenarios and accuracy.
• Validated the method on real-world river and industrial water quality datasets.

Main Results:

• The proposed ensemble model achieved higher accuracy and lower variation in imputed values compared to baseline models.
• Demonstrated superior performance across various missingness scenarios.
• Validated the method's applicability in different hydrological environments.

Conclusions:

• The dynamic weighting ensemble model offers a robust solution for water quality data imputation.
• This approach enhances the reliability of hydrological modeling and data analysis.
• The method shows significant potential for improving water resource management.